Code repository of physical object oriented modeling (POOM) paper

It repositions discrete physical entities (e.g., furniture, air conditioners) as primary setup units, preserving problem integrity while reducing manual intervention. To demonstrate this framework, we implemented a prototype for room-scale indoor airflow simulation with diverse furniture layouts using Python and OpenFOAM.

In addition, we constructed an MCP service based on the POOM framework to enable conversational intelligent simulation powered by LLMs. A front-end and back-end separated interactive web interface was developed to achieve cross-device compatibility, decoupling computational logic from design operations. By adopting the aforementioned workflow, MCP services, web-based interaction and other components, the entire project fully demonstrates the software independence and intuitive operability of the POOM framework.